Multi-to-One Style Transfer: Region-Based Styles Assignment for Single Content Image

Introduction

Multi-to-One Style Transfer is a novel style transfer method, aiming to assemble various styles like puzzle pieces to construct a stylized image 🎨

Requirements

• Python 3.8 or later
• Conda

Usage

1. Clone this repository:

   git clone https://github.com/Charles-T-T/M2O_StyleTransfer.git

2. Create and activate the Conda environment:

   conda env create -f environment.yml
   conda activate StyleTransfer

3. Run the M2O_ST.py :

   python M2O_ST.py

   You can choose your own style images and content images by adding arguments like this:

   python M2O_ST.py --style_paths path/to/style1 path/to/style2 --content_path path/to/content

   Run python M2O_ST.py --help to see all arguments:

   usage: M2O_ST.py [-h] [--style_paths STYLE_PATHS [STYLE_PATHS ...]] [--content_path CONTENT_PATH] [--save_dir SAVE_DIR] [--n_epoch N_EPOCH]
   
   Multi-to-One Style Transfer
   
   optional arguments:
     -h, --help            show this help message and exit
     --style_paths STYLE_PATHS [STYLE_PATHS ...]
                           Paths to style images (provide multiple paths separated by space)
     --content_path CONTENT_PATH
                           Path to the content image
     --save_dir SAVE_DIR   Directory to save the result image
     --n_epoch N_EPOCH     Number of training epochs, default to 1000

Demo

Content	Styles	Result

Acknowledgment

Thanks to Chenjun(a3377596@126.com) for key algorithm design and implementation.